Assessing bilateral differences in paired anthropometric features is an important methodological problem in sports anthropology. The present study included 128 adolescent female athletes (59 rhythmic gymnasts, 58 tennis players, and 11 swimmers). Body composition components were determined using multi-frequency bioelectrical impedance measurements (analyzer InBody 170). Asymmetry coefficients of muscle and fat mass accumulation in the upper (AA) and lower (AL) limbs were calculated using the Nacheva` equation (1986). The percentiles method was applied to distribute the bilaterally studied anthropometric features according to the mean values of the units of asymmetry (UA). Wilkoxon-test was used to assess the statistically significant differences in paired variables. Kruskal-Wallis test was applied to determine the differences in UA between three assessed athlete groups, depending on their age. The differences in body composition components between rhythmic gymnasts (RG), tennis players (TP), and swimmers (SW) were well expressed in all assessed age groups. The most considerable inter-group differences were observed in terms of the asymmetry coefficient in the lean body mass (LBM) with a right direction and body fat mass with a left direction for upper limb fat mass (%, kg), which have signed the highest values in the tennis players group, followed by the RG on the same age. Swimmers had significantly the lowest values of UA for all body segments. A close relation was found between asymmetry in body composition variables and the type of sports activity. Tennis was found as a sport with more pronounced inter-limbs asymmetry.